The Probity-Policy Distinction in the Statistical Evidence Debate

Although the use of statistical testimony at trial has increased dramatically during the past two decades,' tensions between statistics and the law remain. Where the law strives for certainty, statistical evidence frames the world in uncertain, probabilistic terms. This tension has sparked a series of debates in many major law reviews over the value of statistical evidence. Regrettably, these exchanges have not always clarified the central issues. In some cases, the debaters misstate or misunderstand the positions of perceived adversaries. In other cases, the debaters fail to indicate whether their views reflect concerns about the probative efficacy of statistical evidence (either in theory or in practice) or the challenge statistical evidence poses to underlying legal policies and values. Both problems are found in the recent statistical evidence debate published by the Tulane Law Review. The present contribution is divided into three sections. Section I is a selective response to the claims made by Professors Callen, Allen, and Shaviro about an article that Professor Shaviro and I published on overtly probabilistic evidence and methods. Section II attempts to clarify the probity-policy distinction, a distinction that has played a central-albeit hidden role in many of the statistical evidence debates. Section III calls for quantitative training in the legal curriculum and proposes a research agenda that could have important implications for the presentation of quantitative testimony at trial. Section IV is a brief conclusion